Odor-Cued Targeted Reactivation Was Unable to Selectively Benefit Declarative Memories During Sleep

Declarative memories are reactivated-and thereby consolidated-during sleep. Real-life memories are typically nested hierarchically (e.g., memory for making coffee nested within memory for one's morning routine). We tested the specificity of memory reactivation during sleep in humans: is it limited to low-tier items or does it extend to wider contexts? To test this, we adapted a well-replicated design using targeted memory reactivation, which uses non-invasive sensory cues to preferentially reactivate memories during sleep. Thirty-two participants (18 women and 14 men) learned two sets of object locations, each paired with a unique odour. By cueing one odour during non-REM sleep, we tested whether reactivation would benefit the entire learning context or selectively enhance the cued set. Our results show no overall benefit for the cued set over the non-cued one. A more nuanced, encoding-strength-dependent reactivation effect was observed for the cued category relative to the non-cued one. Whereas previous studies showed that odour presentation increased spectral activity in the sigma range, putatively reflecting sleep spindles, we found a sustained (~15 s) inhibition following presentation. The results indicate that cueing did not uniformly benefit the targeted memories. One explanation for these results is that cueing benefits may have generalised across the learning context as a whole rather than impacting a single set of memories. Moreover, our results provide more evidence that initial encoding strength dictates the extent of reactivation effectiveness.